From reading the ACARS data and discussions with others in our unit, it certainly does seem that Cb penetration followed by sensor failures and loss of control is the most likely scenario.
Rather worryingly, from a number of comments, it appears that many pilots do not use their Wx radar properly.
Radar is not simply something that you switch on in the hope that significant returns will pop up - the display must be seen to even test and start using it
Spot on!
The radar detects water droplets only. It does not, indeed
cannot, detect ice. The turbulence mode only detects turbulence that contains water droplets; if there is no rain in the turbulent air, the turbulence remains invisible. On a dark night, if there is no rain in the top of a Cb it is invisible. The Wx radar must be operated in pitch to scan various altitudes ahead: the beam covers a taller column at greater range than at close range and you can pick up the rain that appears lower down in the column that may indicate the presence of a Cb full of ice towering up above it.
Forgive me, a couple of dumb questions for my own edification. Firstly, if we don't know what caused AF to go down until the report comes out, what steps should crew take to avoid the same fate?
Cbs may sprout up in your face very quickly - while working in equatorial climes for decades I have regularly observed Cbs build up from virtually nothing to monsters towering up to 30,000+ feet in the space of fifteen to twenty minutes. This makes it especially important to actively manipulate the Wx radar tilt when flying in known areas of Cb activity and learn to interpret what you see. Do not imagine that what you see is what you get with Wx radar; the truth is, you have to work it and interpret. Honeywell and Rockwell Collins both produce useful pocket sized pilot guides on the subject.
Contact the Flight Ops Support people through these links and since you're pilots, try asking for a freebie:
Honeywell
Rockwell Collins